Image recognition method and apparatus

    公开(公告)号:US11100320B2

    公开(公告)日:2021-08-24

    申请号:US16542597

    申请日:2019-08-16

    Abstract: This disclosure proposes an image recognition method and apparatus. The method comprises: obtaining an image to be recognized; inputting the image into a first preset block of a residual neural network, to obtain a first image feature corresponding to the image; inputting the first image feature into a second preset block of the residual neural network, an attention model, a first convolution layer, and a pooling layer arranged in this order, to obtain first label information corresponding to the image, which has a label correlation representation; inputting the first image feature into a second convolution layer and a bidirectional neural network arranged in this order, to obtain second label information corresponding to the image, which has a label correlation representation; and determining label information corresponding to the image in accordance with the first label information and the second label information.

    Array substrate, display panel, and display device

    公开(公告)号:US11699709B2

    公开(公告)日:2023-07-11

    申请号:US17515246

    申请日:2021-10-29

    CPC classification number: H01L27/1244 G02F1/1339 G02F1/1368 H01L27/1259

    Abstract: The present disclosure provides an array substrate, a display panel, and a display device. The array substrate includes: a substrate; and a first conductive structure, an interlayer insulation layer, and a second conductive structure sequentially disposed on the substrate. The first conductive structure has a first connection portion, the second conductive structure has a second connection portion, and the first connection portion is electrically coupled to the second connection portion through a via penetrating through the interlayer insulation layer. At least one of the first connection portion and the second connection portion is provided with an opening, and an orthographic projection of the opening on the substrate does not overlap an orthographic projection of the via on the substrate.

    Machine Learning Model Training Method and Device and Electronic Equipment

    公开(公告)号:US20230030419A1

    公开(公告)日:2023-02-02

    申请号:US17788608

    申请日:2021-07-05

    Inventor: Tingting Wang

    Abstract: The invention relates to a machine learning model training method and device and electronic equipment, and relates to the technical field of artificial intelligence. The training method includes the following steps: inputting an image sample into a regression machine learning model, extracting a feature map of the image sample by utilizing the regression machine learning model, and determining an identification result of the image sample according to the feature map; inputting the feature map into a classification machine learning model, and determining the membership probability of the image sample belonging to each classification by using the classification machine learning model according to the feature map; calculating a first loss function according to the recognition result and the labeling result of the image sample, and calculating a second loss function according to the membership probability and the labeling result of the image sample; and training a regression machine learning model by using the first loss function and the second loss function.

    Distortion correction method and apparatus, electronic device, and computer-readable storage medium

    公开(公告)号:US11295419B2

    公开(公告)日:2022-04-05

    申请号:US16945336

    申请日:2020-07-31

    Abstract: The present disclosure provides a distortion correction method, a distortion correction apparatus, an electronic device, and a computer-readable storage medium, the distortion correction method including: performing a line detection on the distorted image to obtain multiple straight lines in the distorted image; grouping the straight lines into multiple groups of straight lines; selecting optimal boundary lines of the distorted image from the groups of straight lines; determining multiple initial vertexes of the distorted image according to intersection points between the optimal boundary lines; and performing a perspective transformation processing on the distorted image according to the initial vertexes to obtain a corrected image.

Patent Agency Ranking